1
|
Chen Y, Gao Y, Zhu L, Shao W, Lu Y, Han H, Xie Z. PCNet: Prior Category Network for CT Universal Segmentation Model. IEEE TRANSACTIONS ON MEDICAL IMAGING 2024; 43:3319-3330. [PMID: 38687654 DOI: 10.1109/tmi.2024.3395349] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 05/02/2024]
Abstract
Accurate segmentation of anatomical structures in Computed Tomography (CT) images is crucial for clinical diagnosis, treatment planning, and disease monitoring. The present deep learning segmentation methods are hindered by factors such as data scale and model size. Inspired by how doctors identify tissues, we propose a novel approach, the Prior Category Network (PCNet), that boosts segmentation performance by leveraging prior knowledge between different categories of anatomical structures. Our PCNet comprises three key components: prior category prompt (PCP), hierarchy category system (HCS), and hierarchy category loss (HCL). PCP utilizes Contrastive Language-Image Pretraining (CLIP), along with attention modules, to systematically define the relationships between anatomical categories as identified by clinicians. HCS guides the segmentation model in distinguishing between specific organs, anatomical structures, and functional systems through hierarchical relationships. HCL serves as a consistency constraint, fortifying the directional guidance provided by HCS to enhance the segmentation model's accuracy and robustness. We conducted extensive experiments to validate the effectiveness of our approach, and the results indicate that PCNet can generate a high-performance, universal model for CT segmentation. The PCNet framework also demonstrates a significant transferability on multiple downstream tasks. The ablation experiments show that the methodology employed in constructing the HCS is of critical importance. The prompt and HCS can be accessed at https://github.com/PKU-MIPET/PCNet.
Collapse
|
2
|
Gao C, Wu X, Cheng X, Madsen KH, Chu C, Yang Z, Fan L. Individualized brain mapping for navigated neuromodulation. Chin Med J (Engl) 2024; 137:508-523. [PMID: 38269482 PMCID: PMC10932519 DOI: 10.1097/cm9.0000000000002979] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/24/2023] [Indexed: 01/26/2024] Open
Abstract
ABSTRACT The brain is a complex organ that requires precise mapping to understand its structure and function. Brain atlases provide a powerful tool for studying brain circuits, discovering biological markers for early diagnosis, and developing personalized treatments for neuropsychiatric disorders. Neuromodulation techniques, such as transcranial magnetic stimulation and deep brain stimulation, have revolutionized clinical therapies for neuropsychiatric disorders. However, the lack of fine-scale brain atlases limits the precision and effectiveness of these techniques. Advances in neuroimaging and machine learning techniques have led to the emergence of stereotactic-assisted neurosurgery and navigation systems. Still, the individual variability among patients and the diversity of brain diseases make it necessary to develop personalized solutions. The article provides an overview of recent advances in individualized brain mapping and navigated neuromodulation and discusses the methodological profiles, advantages, disadvantages, and future trends of these techniques. The article concludes by posing open questions about the future development of individualized brain mapping and navigated neuromodulation.
Collapse
Affiliation(s)
- Chaohong Gao
- Sino–Danish College, University of Chinese Academy of Sciences, Beijing 100190, China
| | - Xia Wu
- Brainnetome Center, National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China
| | - Xinle Cheng
- School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing 100190, China
| | - Kristoffer Hougaard Madsen
- Department of Applied Mathematics and Computer Science, Technical University of Denmark, Kongens Lyngby 2800, Denmark
- Danish Research Centre for Magnetic Resonance, Centre for Functional and Diagnostic Imaging and Research, Copenhagen University Hospital Amager and Hvidovre, Hvidovre 2650, Denmark
| | - Congying Chu
- Brainnetome Center, National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China
| | - Zhengyi Yang
- Brainnetome Center, National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China
| | - Lingzhong Fan
- Sino–Danish College, University of Chinese Academy of Sciences, Beijing 100190, China
- Brainnetome Center, National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China
- School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing 100190, China
- CAS Center for Excellence in Brain Science and Intelligence Technology, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China
- School of Health and Life Sciences, University of Health and Rehabilitation Sciences, Qingdao, Shandong 266000, China
| |
Collapse
|
3
|
Liu H, Nie D, Yang J, Wang J, Tang Z. A New Multi-Atlas Based Deep Learning Segmentation Framework With Differentiable Atlas Feature Warping. IEEE J Biomed Health Inform 2024; 28:1484-1493. [PMID: 38113158 DOI: 10.1109/jbhi.2023.3344646] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/21/2023]
Abstract
Deep learning based multi-atlas segmentation (DL-MA) has achieved the state-of-the-art performance in many medical image segmentation tasks, e.g., brain parcellation. In DL-MA methods, atlas-target correspondence is the key for accurate segmentation. In most existing DL-MA methods, such correspondence is usually established using traditional or deep learning based registration methods at image level with no further feature level adaption. This could cause possible atlas-target feature inconsistency. As a result, the information from atlases often has limited positive and even counteractive impact on the final segmentation results. To tackle this issue, in this paper, we propose a new DL-MA framework, where a novel differentiable atlas feature warping module with a new smooth regularization term is presented to establish feature level atlas-target correspondence. Comparing with the existing DL-MA methods, in our framework, atlas features containing anatomical prior knowledge are more relevant to the target image feature, leading the final segmentation results to a high accuracy level. We evaluate our framework in the context of brain parcellation using two public MR brain image datasets: LPBA40 and NIREP-NA0. The experimental results demonstrate that our framework outperforms both traditional multi-atlas segmentation (MAS) and state-of-the-art DL-MA methods with statistical significance. Further ablation studies confirm the effectiveness of the proposed differentiable atlas feature warping module.
Collapse
|